Explain the advantage of creating a linear function from a set of data; why bother to do this?

If the data you list has a linear relationship, then creating a linear function from the data might be helpful because it would define the relationship and help you to answer questions related to it that the data you have at present might not.

Here is an example. Height and shoe size have a linear relationship. Let's say you plot the point (height in inches, shoe size in US women's sizes) for 50 people or so. You would see a strong linear relationship. The points wouldn't make a perfect line but very close to it. As height increases so does shoe size and it does so in a linear way. Now if we call the height h and the shoe size s then we would have an equation that relates h and s. So that if we wanted to know the shoe size of someone 96 inches (8 feet high) tall we could get that (or very close to it) from the equation even if there was no one in our sample that was that tall. Similarly we could solve for height if we knew shoe size.